Timeline and text
Edit audio through a transcript
Transcribe speech, select audio by words, and remove filler words in one undoable edit.
Create a transcript
Choose View → Transcript… to open the transcript panel. Pick a model, either Whisper or Grok (xAI), both provided through Puter and both returning word timings, and choose one of 19 languages or Auto-detect. The Transcribe button shows how much audio will be sent. Transcription is an online action that can ask you to sign in to Puter.
Check names, punctuation, and word timing against the recording. Audio you add later appears as untranscribed blocks with a Transcribe everything untranscribed button. If you change the model or language after transcribing, a Re-transcribe bar appears so you can redo those takes with the new settings, or leave them as they are.
Select and delete words
Click a word to move the cursor to it; the editor nudges the position to the nearest audible onset so playback starts cleanly. Double-click to select a word, drag or Shift-click to select more, and use Shift with the arrow keys to extend the selection word by word. The corresponding time span is selected on the waveform, so you can listen, preview the cut, or apply an effect.
Press Backspace or Delete to remove the selected words. The cut edges settle into nearby pauses when the audio has them, so joins fall between words rather than inside them. Escape clears the selection, and undo restores the edit. Words follow splits, merges, trims, and cuts without retranscription. This edits audio through recognised words; typing new text does not create speech.
Review filler words
The panel counts supported fillers such as “um” and “uh” for the transcript’s language. Review steps through them one at a time with Prev, Remove, and Skip, selecting and revealing each on the waveform so you can listen first. Remove all deletes every filler in one undoable step. Filler matching depends on the language and recording.
While the transcript panel is open, Paste and Reverse are unavailable because they would rebuild the audio behind the words; choose View → Close transcript first. View → Clear transcript discards the text. Save Project stores transcripts with the audio, and Voice Recorder imports can include an existing word-level transcript.